多源遥感融合技术在古湖岸识别中的应用 ——以察尔汗盐湖为例

Demonstration of multi-source remote sensing fusion technology to identify ancient lake shorelines - Take the Qarhan salt lake as an example

中文关键词:  SAR  多源遥感  数据融合技术  察尔汗盐湖  古湖岸线

英文关键词:SAR  Multi-source remote sensing  Data fusion technology  Qarhan Salt Lake  Ancient lake shoreline

中文摘要:

      多源遥感融合技术在识别被侵蚀或被覆盖的古湖岸线方面具有独特的优势。察尔汗盐湖为我国最大的第四纪液体钾盐矿床，识别该盐湖晚更新世以来的多级古湖岸及其湖面变化过程，对于保障我国的钾肥安全具有重要的意义。本研究以合成孔径雷达（SAR）哨兵一号数据（Sentinel-1A）为数据源，应用IHS融合、PC融合、GS融合和HSV融合4种融合方法对高分辨率多光谱哨兵二号数据（Sentinel-2B）在察尔汗盐湖区域进行了融合实验。并对融合结果进行精度评价，选取信息熵、互信息、平均梯度、空间频率、标准差、相关系数、差异相关性总和以及视觉保真度8种客观评价指标对融合结果进行了计算与分析。利用野外实测古湖岸的遗迹高程结合数字高程数据（DEM）对岸线进行模拟，并生成地形剖面图，来恢复晚更新世以来察尔汗盐湖古湖岸动态变化。结果表明：融合方法中的HSV算法是最适合应用于察尔汗盐湖古湖岸线的提取，对恢复晚更新世以来被侵蚀被覆盖的三级古湖岸线地貌，具有较好的效果。

英文摘要:

      Multi-source remote sensing fusion technology has a unique advantage in identifying the eroded or covered ancient lake shoreline. Qarhan Salt Lake is the largest quaternary liquid potassium salt deposit in China. It is of great significance to identify the multistage paleoshoreland and its lake surface variation process since the late Pleistocene in Qarhan Salt Lake to ensure the safety of potash fertilizer in China. In this study, the Sentinel-1A data of synthetic aperture radar (SAR) was used as the data source, and four fusion methods, namely IHS fusion, PC fusion, GS fusion and HSV fusion, were used to fusion high-resolution multi-spectral Sentinel-2B data in the Qarhan Salt Lake region. The accuracy of the fusion results was evaluated, and 8 objective evaluation indexes including information entropy, mutual information, average gradient, spatial frequency, standard deviation, correlation coefficient, differential correlation sum and visual fidelity were selected to calculate and analyze the fusion results. The elevation of the ancient lake shore measured in the field combined with digital elevation data (DEM) was used to simulate the shoreline, and the topographic profile map was generated to restore the dynamic changes of the ancient shore of the Qarhan salt lake since the Late Pleistocene.The results show that the HSV algorithm in the fusion method is the most suitable for the extraction of the ancient shoreline of Qarhan Salt Lake, and has a good effect on restoring the third-order ancient shoreline landforms that have been eroded and covered since the late Pleistocene.